1: \documentclass[aps,pre,twocolumn,amsfonts,amssymb,amsmath,showpacs]{revtex4}
2: \usepackage[dvips]{graphicx}
3:
4: \addtolength{\topmargin}{10mm}
5:
6: %%%%%%%%%%%%%%%%%%%%%%%%%%
7: % KEYSTROKE SAVING MACROS
8: %\input{abbr.tex}
9: % Text
10: \newcommand{\eg}[1]{{\it e.g.\/}\ifx#1.\else\expandafter#1\fi}
11: \newcommand{\ie}[1]{{\it i.e.\/}\ifx#1.\else\expandafter#1\fi}
12:
13: % Generic maths notations
14: \renewcommand{\@}{\partial}
15: \newcommand{\const}{\mathrm{const}}
16: \newcommand{\Df}[2]{\displaystyle{\frac{\d #1}{\d #2}}}
17: \renewcommand{\d}{\mathrm{d}}
18: \newcommand{\df}[2]{\frac{\partial #1}{\partial #2}}
19: \newcommand{\ddf}[2]{\frac{\partial^2 #1}{\partial #2^2}}
20: %\newcommand{\det}{\mathrm{det}}
21: \newcommand{\eq}[1]{\eqref{#1}}
22: \newcommand{\Mx}[1]{\left[\begin{array}{cc}#1\end{array}\right]}
23: \newcommand{\mx}[1]{\mathbf{#1}}
24: \newcommand{\Real}{\mathbb{R}}
25: \renewcommand{\Re}{\mathrm{Re}}
26:
27: % Specific notations
28: \newcommand{\D}{\mx{D}} % diffusion matrix
29: \newcommand{\Dr}{\D_r} % diffusion matrix at the coexistence eq
30: \newcommand{\DP}{D_{\P}} % prey diffusivity
31: \newcommand{\DZ}{D_{\Z}} % predator diffusivity
32: \newcommand{\dt}{\Delta t} % time discretization step
33: \newcommand{\dx}{\Delta x} % space discretization step
34: \newcommand{\F}{\mx{F}} % jacobian mx of nonlinear kinetics
35: \newcommand{\Fr}{\F_r} % jacobian mx at the coexistence eq
36: \newcommand{\f}{\mx{f}} % nonlinear kinetics
37: \newcommand{\hm}{h_{-}} % evasion coefficient
38: \newcommand{\hp}{h_{+}} % pursuit coefficient
39: \newcommand{\I}{\mx{I}} % identity mx
40: \renewcommand{\k}{k} % wavenumber
41: \newcommand{\lyap}{\lambda} % growth rate for perturbations of given wavenumber
42: \renewcommand{\u}{\mx{u}} % dynamic vector field
43: \newcommand{\uo}{\mx{u}_o} % the limit cycle
44: \newcommand{\ur}{\mx{u}_r} % the resting state aka coexistence eq
45: \renewcommand{\v}{\mx{v}} % small perturbation
46: \renewcommand{\P}{u} % prey density
47: \renewcommand{\Pr}{\P_r} % prey density at coexistence eq
48: \newcommand{\Pc}{\P_*} % control value of prey density
49: \newcommand{\Z}{v} % predator density
50: \newcommand{\Zr}{\Z_r} % predator density at coexistence eq
51:
52: % Handling figures
53: \newcommand{\Fig}[1]{Fig.~\ref{fig:#1}}
54: \newcommand{\Figs}[1]{Figs.~\ref{fig:#1}}
55: \newcommand{\fig}[1]{fig.~\ref{fig:#1}}
56: \newcommand{\figref}[1]{\ref{fig:#1}}
57: % Wide figure, across the page
58: \newcommand{\dblfigure}[3]{
59: \begin{figure*}[tb!]
60: \includegraphics{#1}
61: \caption[]{#2}
62: \label{fig:#3}
63: \end{figure*}
64: }
65: % Narrow figure, within the column
66: \newcommand{\sglfigure}[3]{
67: \begin{figure}[tb!]
68: \includegraphics{#1}
69: \caption[]{#2}
70: \label{fig:#3}
71: \end{figure}
72: }
73:
74:
75: %%%%%%%%%%%%%%%%%%%%%%%%%%
76: \begin{document}
77: \title{Spontaneous traveling waves in oscillatory systems with cross diffusion}
78: \author{V. N. Biktashev}
79: \affiliation{Department of Mathematical Sciences,
80: University of Liverpool, Liverpool L69 7ZL, UK }
81: \author{M. A. Tsyganov}
82: \affiliation{
83: Institute of Theoretical and Experimental Biophysics,
84: Pushchino, Moscow Region, 142290, Russia}
85: \date{\today}
86: \begin{abstract}
87: We identify a new type of pattern formation in spatially distributed
88: active systems. We simulate one-dimensional two-component systems
89: with predator-prey local interaction and pursuit-evasion taxis
90: between the components. In a sufficiently large domain, spatially
91: uniform oscillations in such systems are unstable with respect to
92: small perturbations. This instability, through a transient regime
93: appearing as spontanous focal sources, leads to establishment of
94: periodic traveling waves. The traveling waves regime is
95: established even if boundary conditions do not favor such solutions.
96: The stable wavelength are within a range bounded both from above and
97: from below, and this range does not coincide with instability bands
98: of the spatially uniform oscillations.
99: \end{abstract}
100: \pacs{%
101: 87.10.+e% Biol and med physics: General theory and mathematical aspects
102: , 02.90.+p% Other topics in mathematical methods in physics
103: }
104: \maketitle
105:
106: \section{Introduction}
107:
108: Dissipative structures, \ie\ patterns in spatially extended
109: systems away from equilibrium have been intensively studied for many
110: decades. A very comprehensive review can be found in
111: \textcite{Cross-Hohenberg-1993}; results obtained since then would
112: probably require an even more extensive review. A very popular
113: class of mathematical models is the reaction-diffusion systems with
114: diagonal diffusion matrices. There have been numerous indications
115: that non-diagonal elements in diffusion matrices, \ie\
116: cross-diffusion, can lead to new nontrivial effects not observed in
117: classical reaction-diffusion systems, \eg\ \emph{quasi-solitons} in
118: systems with excitable reaction part
119: \cite{QS1,QS2,QS3,QS5,QS6}. However oscillatory systems are more prevalent
120: than excitable and nontrivial effects of cross-diffusion in oscillatory systems have not been
121: studied yet. Here we consider an example where the
122: reaction part of the system is dissipative while the diffusion part
123: is not. We describe spontaneously generated periodic waves, and
124: identify the features of these waves that indicate that we are
125: dealing here with a phenomenon not seen before.
126:
127: A general formulation of a reaction-diffusion system with nonlinear diffusion is
128: %
129: \begin{equation}
130: \df{\u}{t} = \f(\u) + \nabla(\D(\u)\nabla\u), \qquad
131: \u,\f\in\Real^N, \; \D\in\Real^{N\times N}. \label{RD}
132: \end{equation}
133: %
134: Both the reaction term $\f(\u)$ and the diffusion term $\nabla(\D(\u)\nabla\u)$
135: in the right-hand side represent dissipative processes, which for
136: diffusion implies that matrix $\D\in\Real^{N\times N}$
137: is positive (semi-)definite, typically diagonal
138: with non-negative elements. A huge amount of results have been
139: obtained about pattern formation described by such models. However,
140: many physical situations lead to non-diagonal elements in $\D$, \ie\
141: cross-diffusion (see \eg\ discussions in \cite{UFN07,Vanag-Epstein-2009}).
142: Some such situations may be adequately
143: described by $\D$ whose eigenvalues have zero real part, \eg\ when the
144: self-diffusion of components is negligible. In such cases reaction
145: part is dissipative and the ``diffusion'' part is not. Physical
146: consequences of such ambivalence are little understood yet.
147:
148: Cross diffusion has been seen to produce interesting phenomena, such
149: as fronts, pulses and stationary periodic structures (see \eg\
150: \cite{del-Castillo-Negrete-etal-2002,Chung-Peacock-2007}
151: among many other works), however phenomenologically similar regimes
152: are known in reaction-diffusion systems, too.
153:
154: In a recent series of works we have described unusual phenomena, such
155: as quasi-solitons and their variations, in excitable systems in which
156: linear or nonlinear cross-diffusion was added to or replaced
157: self-diffusion (see \eg\ \cite{QS1,QS2,QS3,QS5,QS6}). The ability of a medium
158: to conduct solitary waves is stipulated by its excitable kinetics
159: described by the reaction term $\f(\u)$, whereas specifics of their
160: interaction are also due to the cross-diffusion terms. However,
161: excitability is a relatively exotic, albeit very important, type of
162: behaviour compared to oscillations. For instance, in population
163: dynamics, plausible excitable predator-prey models have been proposed
164: \cite{Truscott-Brindley-1994} but we are not aware of reliable observations
165: of natural systems described by such models. On the other hand,
166: oscillatory behaviour in predator-prey systems is textbook material
167: \cite{Murray-2002,Britton-2003} and there are plentiful observational
168: data on traveling waves in cyclic populations
169: \cite{Sherratt-Smith-2008}.
170:
171: Solitary waves in oscillatory systems are not feasible, and it is not
172: clear what new features cross-diffusion may impose.
173:
174: %%%%%%%%%%%%%%%%%%%%%%%%
175: % \input{density.tex}
176: \dblfigure{fig1}{
177: Different regimes resulting from random perturbaton of uniform
178: oscillations.
179: Shown are density plots: space $x$ is horizontal, time $t$ is
180: vertical increasing upwards, $\P=1$ corresponds to black and $\P=0$
181: corresponds to white.
182: (a) TB model, $L=15$, taxis ($\hm=1$, $\hp=0$, $\DP=\DZ=0$), periodic boundary conditions, $t\in[0,2500]$.
183: (b) Same, no-flux boundary conditions, $t\in[0,1200]\cup[43800,45000]$.
184: (c) Same as (a) except $\DP=\DZ=0.05$, $\hp=\hm=0$.
185: %
186: (d) Same as (a) except $w=0.07$, $L=10$, $t\in[0,1200]\cup[36300,37500]$.
187: % p5gmw1-stackcut.pgm: p5gmw1-1000-1240cut.pgm p5gmw1-insert.pbm p5gmw1-0-240cut.pgm
188: (e) Same as (a) except $\hp=0.1$, $L=50$, $t\in[0,1200]\cup[5000,6200]$.
189: % 1e-D0p02-5000-stackcut.pgm: 1e-D0p02-5000-380-780cut.pgm 1e-D0p02-5000-insert.pbm 1e-D0p02-5000-0-80cut.pgm
190: (f) Same as (a) except $\hp=0.1$, $\DP=\DZ=0.02$, $L=50$, $t\in[0,400]\cup[1900,3900]$.
191: (g) RM model, $\hm=1$, $\hp=\DP=\DZ=0$, $L=25$, $t\in[0,2500]$.
192: }{density}
193: %%%%%%%%%%%%%%%%%%%%%%%%
194:
195: The purpose of this communication is to describe new phenomena we have
196: observed in oscillatory systems with ``pursuit-evasion'' nonlinear
197: cross-diffusion interaction between the components.
198:
199: \section{The models}
200:
201: % We restrict ourselves to
202: % purely empirical observations by numerical simulations; any
203: % theoretical analysis is deferred to other publications.
204:
205: We consider two predator-prey models with cross-diffusion terms of
206: ``pursuit-evasion'' mutual taxis,
207: %
208: \begin{eqnarray}
209: \df{\P}{t} &=& f(\P,\Z) + \DP \ddf{\P}{x} + \hm \df{}{x}\left(\P\df{\Z}{x}\right) , \nonumber\\
210: \df{\Z}{t} &=& g(\P,\Z) + \DZ \ddf{\Z}{x} - \hp \df{}{x}\left(\Z\df{\P}{x}\right) , \label{RT}
211: \end{eqnarray}
212: %
213: for $(x,t)\in[0,L]\times[0,t_{\max}]$
214: for two reaction kinetics, the Truscott-Brindley (TB) model
215: \cite{Truscott-Brindley-1994}
216: %
217: \begin{eqnarray}
218: f(\P,\Z) &=& \beta \P(1-\P) - \Z\P^2/(\P^2+\nu^2), \nonumber \\
219: g(\P,\Z) &=& \gamma \Z\P^2/(\P^2+\nu^2) - w\Z , \label{TB}
220: \end{eqnarray}
221: %
222: where $\beta=0.43$, $\nu=0.053$, $\gamma=0.1$ and $w=0.055$ unless
223: stated otherwise, and the
224: Rosenzweig-MacArthur (MA) model
225: \cite{Rosenzweig-MacArthur-1963,Britton-2003,Sherratt-Smith-2008}
226: %
227: % \nu=k
228: % \gamma=\sigma
229: % w=\mu
230: \begin{eqnarray}
231: f(\P,\Z) &=& \beta \P(1-\P) - \Z\P/(\P+\nu), \nonumber \\
232: g(\P,\Z) &=& \gamma \Z\P/(\P+\nu) - w\Z , \label{RM}
233: \end{eqnarray}
234: %
235: where $\beta=1$, $\nu=0.3$, $\gamma=0.15$ and $w=0.03$ unless stated
236: otherwise.
237: Here $\P$ represents
238: prey, $\Z$ predators, the term with $\hp$ describes pursuit of prey by
239: predators and the term with $\hm$ describes evasion of predators by prey.
240: The simulation were done on an interval $x\in[0,L]$ with
241: periodic or Neumann boundary conditions for both
242: components, using forward Euler stepping in time,
243: center differences for the diffusion terms and upwind difference for
244: the taxis terms, see \textcite{QS2} for details and justification.
245: Except where stated otherwise, we used discretization
246: steps $\dx=0.1$ and $\dt=4\cdot10^{-4}$.
247:
248: \section{Numerical observations}
249:
250: %%%%%%%%%%%%%%%%%%%%%%%%
251: % \input{profs.tex}
252: \sglfigure{fig2}{ (color online)
253: Examples of profiles of spontaneously established periodic
254: waves.
255: Shown are dependencies of $\P$ and $\Z$ on $x$ at fixed $t$,
256: and direction of propagation by arrows.
257: Parameters are the same as in \fig{density} except
258: interval lentgh $L$, specifically,
259: (a) as in \fig{density}(a),
260: (b) as in \fig{density}(d),
261: (c) as in \fig{density}(e),
262: (d) as in \fig{density}(f) and
263: (e) as in \fig{density}(g).
264: }{profs}
265: %%%%%%%%%%%%%%%%%%%%%%%%
266:
267: \Figs{density} and \ref{fig:profs} illustrate the phenomenon of the
268: spontaneous onset of periodic waves. Starting from arbitrary spatially-uniform
269: intial conditions at $t=0$, after a transient allowed to establish
270: uniform oscillations, perturbations were introduced and subsequent
271: evolution observed. The perturbation was introduced at half of the
272: grid points chosen randomly, where at $t=300$ the values of $\P$ were
273: replaced by randomly chosen numbers in the interval between $0.15$ and
274: $0.45$. \Fig{density} shows space-time density plots and \fig{profs} illustrate
275: selected profiles of the emerging wavetrains.
276:
277: In the TB model with periodic boundary conditions (\fig{density}(a)),
278: after a ``random'' transient lasting two or three bulk oscillation
279: periods, patterns start to emerge: waves start ``from nowhere'' and
280: annihilate upon collision with other such waves. After a few periods
281: of such collisions, the waves propagating leftwards win over and a
282: periodic wavetrain establishes which then persists. Different seeds
283: in the random number generator produce solutions differing in details
284: but always leading to periodic trains, leftward and rightward
285: propagating with equal probability (compare
286: density plots and wave profiles in \fig{density} and \fig{profs},
287: which corresponded to different simulations with the same parameter sets).
288:
289: Impenetrable boundaries do not allow periodic wavetrain solutions;
290: however the tendency to establish periodic wavetrains is observed even
291: then. In \fig{density}(b) rightward propagating waves win over. Their impact
292: with the right boundary $x=L$ is with partial reflection, when the
293: reflected wave is weak and soon decays; note that this behaviour is
294: typical for collision of solitary excitation waves in such systems
295: \cite{QS2}. The left boundary has a quenching effect, but at a
296: distance from it waves emerge spontaneously. This distance varies
297: irregularly, indicating that spontaneous generation of waves is
298: associated with an instability, thus sensitive dependence on initial
299: conditions and probably chaotic dynamics. This irregular pattern
300: persists for a long time.
301:
302: This behaviour is in a contrast with a system with the same
303: kinetics but pure diffusional spatial terms: in \fig{density}(c), similar
304: initial random perturbations lead very quickly to re-establishment of
305: spatially uniform oscillations.
306:
307: The parameters used in \fig{density}(a) are close to the boundary of
308: the oscillatory regime in the TB model (achieved \eg\ at
309: $w\approx0.053$ with other parameters fixed). When parameters are
310: further into the oscillatory region, spontaneous generation of
311: periodic wavetrains is still observed, although the transient period
312: of spontaneous wavelet generations and collisions lasts longer, see
313: \fig{density}(d).
314:
315: We have also found that prevalence of the ``evasion'' taxis ($\hm$ coefficient) helps
316: generation of periodic trains, but $\hp=0$ is not necessary, and such
317: generation can be observed with the ``pursuit'' taxis present as well,
318: see \fig{density}(e).
319:
320: Spontaneous generation of periodic trains is observed in the RM model
321: as well, see \fig{density}(f).
322:
323: %%%%%%%%%%%%%%%%%%%%%%%%
324: % \input{lchange.tex}
325: \sglfigure{fig3}{ (color online)
326: Variability of the wavetrains at changing $L$.
327: (a) Number of waves in the interval $[0,L]$ as $L$
328: gradually increases (by $0.2$ every 2000 time units, solid red
329: line) and decreases (by $0.2$ every 1000 time units, dashed blue
330: line). Oblique dashed lines: $n=L/2.5$ and $n=L/8$ to guide the eye.
331: (b) Wave periods measured at a point, as function of
332: $L$ as it decreases.
333: (c) Same, as $L$ increases.
334: (d) Density plots of two episodes of simulation
335: of 1200 time units duration each,
336: with $L$ increasing
337: by $0.2$ every 1000 time units.
338: Lower episode: soft transition from steady 1-wave
339: solution to modulated 1-wave solution ($L: 4.8 \to 5.6$).
340: Upper episode: subsequent
341: sudden transition from modulated 1-wave solution to
342: a steady 2-wave solution ($L:7.2 \to 7.6$).
343: }{lchange}
344: %%%%%%%%%%%%%%%%%%%%%%%%
345:
346: The spontaneously emerging periodic wavetrains typically had
347: wavelengths in a limited range. To check whether this is dictated by
348: initial conditions or is due to limitations of the system, we
349: performed simulations in a circle, \ie\ an interval with periodic
350: boundary conditions, of a slowly changing length $L$. We started from
351: an established propagating wave in a circle. Then we changed the
352: length $L$ of the circle in small steps, allowing sufficient time
353: between the steps for the waves to adjust. During the simulation we
354: monitored the number of waves $n$, determined via the number of points
355: where $\P$ crossed the level $\P=\Pc=0.2$, and the periods $T$ defined as
356: intervals between $\P$ crossing the level $\P=\Pc$ in the positive
357: direction. Results of one such simulation are shown in \fig{lchange}.
358:
359: The number of waves $n$ in the interval did not remain constant
360: (\fig{lchange}(a)), but spontaneously adjusted so as to keep the
361: average wavelength within certain limits: between approximately 2.5 and 8 in the
362: simulation shown. This number was not a unique function of the
363: interval length: changing $L$ upwards and downwards produced different
364: dependencies $n(L)$, \ie\ we have hysteresis. Simulations at slower
365: rate of change of $L$ slightly changed the $n(L)$ dependencies but the
366: hysteresis stayed. Near the transition points where $n$ changed the
367: value, the propagation of the waves was non-stationary, and was always
368: for $L$ just below the transitional value, whether it was decreasing
369: (\fig{lchange}(b)) or increasing (\fig{lchange}(c)). Increasing $L$
370: had a noticeably more destabilizing effect than decreasing.
371:
372: The nature of the non-stationary solutions is illustrated by the
373: density plots shown in \fig{lchange}(d). Starting from an $n=1$
374: solution, an increase of $L$ above the value of $L\approx5$ leads to
375: an instability of the steady propagating wave solution. This is a
376: soft, Eckhaus-type
377: instability and leads to a mild modulation of the wave, producing
378: a seemingly two-periodic motion. The amplitude of the modulations
379: grows as $L$ increases, until at $L=7.6$ a qualitative transformation
380: occurs. A gap between the wave and its own copy around the circle
381: grows so big that at a certain moment it is sufficient to allow
382: spontaneous generation of another wave, leading to an $n=2$
383: solution. This solution is steady, \ie\ propagates without
384: modulations, until $L$ grows so big it in turn becomes unstable etc.
385:
386: \section{Preliminary theoretical considerations}
387:
388: Substantial theoretical analysis of the phenomenon of the spontaneous
389: traveling periodic waves is beyond the scope of this communication.
390: Here we consider one naive approach and then some known pattern
391: formation mechanisms, which a priori might look relevant to this
392: phenomenon, only to eliminate them, as not providing a satisfactory
393: explanation. We will refer to the
394: historical review by
395: \textcite[p.~870]{Cross-Hohenberg-1993} (CH for brevity), and to
396: a recent symmetry based classification of instabilities and
397: bifurcations of periodic dissipative waves and structures given by
398: \textcite[p.~2680]{Rademacher-Scheel-2007} (RS for
399: brevity).
400:
401:
402: \paragraph{It is not captured by lambda-omega approach}
403:
404: The simple class of two-component reaction-diffusion systems
405: introduced by \textcite{Kopell-Howard-1973} and called
406: ``lambda-omega systems'', and closely related to the complex Ginzburg-Landau
407: equation, allows exact solutions in the form of periodic waves. It has
408: offered qualitative insight in many nonlinear wave phenomena,
409: including periodic waves in cyclic
410: populations~\cite{Sherratt-Smith-2008}. However, it does not seem to be
411: helpful in our present case. The modification of the lambda-omega
412: system, corresponding to the choice of signs of taxis terms in \eq{RT}
413: is
414: %
415: \begin{equation}
416: \df{z}{t} = (\Lambda(|z|) + i\Omega(|z|)) z - i\nabla^2z
417: \end{equation}
418: %
419: where $z$ is a complex dynamic variable representing $\P+i\Z$,
420: and the purely imaginary diffusivity here corresponds to the absence of
421: self-diffusion, $\DP=\DZ=0$. Then the periodic traveling wave ansatz
422: $z=a\,\exp[i(\omega t-\k x)]$, $a,\omega,\k\in\Real$, gives the finite system
423: %
424: \[ \Lambda(a) =0 , \qquad \omega = \Omega(a) - \k^2, \]
425: %
426: \ie\ all waves have the same amplitude which is a root of $\Lambda()$,
427: and exist for all wavelengths $\k$ rather than in a finite interval.
428: Stability analysis and consideration of nonzero self-diffusion do
429: not help either.
430:
431: \paragraph{It does not emerge via Turing mechanism.} The instability of
432: spatially-uniform solutions in favour of non-oscillatory,
433: spatially-periodic solutions with periods in a finite range is, of
434: course, a defining feature of the Turing patterns, called just so by
435: RS and classified as type $I_s$ in CH nomenclature. Cross-diffusion
436: can provide an alternative to the original Turing's short range
437: inhibition - long range activation condition. Indeed Turing-type
438: instabilities and spontaneously occurring, self-supporting
439: time-stationary spatially periodic patterns have been observed in
440: locally multistable systems with cross-diffusion
441: \cite{del-Castillo-Negrete-etal-2002}. Our present observations are
442: different in that here we are dealing with time-oscillating phenomena
443: not just space-oscillating.
444:
445: \paragraph{It does not emerge via Turing-Hopf mechanism.}
446:
447: A Hopf bifurcation of the spatially uniform equilibrium at a
448: nonzero wavelength, is called ``Hopf'', ``oscillatory Turing'' and
449: ``Turing Hopf'' instability by RS, classifed as type $I_o$ in CH
450: nomenclature, and also known as short-wave instability or
451: finite-wavelength instability. It can lead to stable periodic
452: propagating waves, in lasers, fluid convection and reaction-diffusion
453: models
454: \cite{Swift-Hohenberg-1977,Haken-1983,Livshits-1983,Lega-etal-1994}.
455: In reaction-diffusion context, such waves have been observed
456: experimentally and in simulations in populations and BZ reaction
457: \cite{Mendelson-Lega-1998,Vanag-Epstein-2002}. However, the standard
458: way such instability occurs in systems \eq{RD} implies existence of an
459: equilibrium that is stable with respect to spatially uniform
460: perturbations, which we do not have here, and it only can occur if
461: $N\ge3$ whereas we have only two components, $\P$ and $\Z$.
462:
463: Specifically, for $\u(x,t)=\ur+\v\,e^{\lambda t+i\k x}$ where
464: $\ur=(\Pr,\Zr)$ is the spatially uniform equilibrium and $|\v|\ll1$,
465: we have the characteristic equation
466: %
467: \[
468: \det\left( \Fr - \Dr\k^2 - \lambda\I \right) = 0,
469: \]
470: %
471: where
472: $\Fr=\F(\ur)=\left(\@\f/\@\u\right)_{\u=\ur}=\Mx{f_{11}&f_{12}\\f_{21}&f_{22}}$
473: is the Jacobian matrix of the reaction terms and $\Dr=\D(\ur)$ is the
474: diffusion matrix, both evaluated at the equilibrium. Considering for
475: simplicity the cases of \fig{density}(a,b,d,f) where $\Dr=\Mx{0 & \hm\Pr \\ 0 &
476: 0}$, we have
477: %
478: \[
479: \lambda=\frac12\left( f_{11} + f_{22} \pm \sqrt{
480: (f_{11}-f_{22})^2+4f_{12}f_{21} - f_{21} \hm\Pr\k^2
481: }\right)
482: \]
483: %
484: which for
485: $\k^2>\max\left( ((f_{11}-f_{22})^2+4f_{12}f_{21})/(\hm\Pr) \,,\, 0\right)$
486: gives oscillatory behaviour of perturbations, but then
487: $\Re(\lambda)=(f_{11}+f_{22})/2=\const$ whereas it has to have a
488: maximum at a positive $\k^2$ for this mechanism to be relevant.
489:
490: \paragraph{It does not emerge via Turing-Hopf instability of spatially uniform oscillations}
491:
492: %%%%%%%%%%%%%%%%%%%%%%%%
493: % \input{instab.tex}
494: \sglfigure{fig4}{ (color online)
495: %
496: Emergence of spontaneous periodic waves through instabilities.
497: %
498: (a) Spectrum $\lyap(\k)$ of harmonic perturbations of the spatially uniform
499: oscillations (black lines with points), and the histograms of the
500: wavenumbers of spontaneous wavetrains in the simulations shown in
501: \fig{lchange} for increasing $L$ (red solid line) and decreasing $L$
502: (blue dashed line).
503: %
504: (b) Same, for the RM model, parameters as in \fig{density}(f),
505: histograms obtained by increasing
506: $L$ by 0.2 every 2000 time units from 5 to 50 and
507: decreasing it back to 2 by 0.2 every 1000 time units.
508: %
509: (c) Emergence of standing periodic waves via an instability
510: of the spatially uniform oscillations. Parameters as in \fig{density}(a)
511: and \fig{lchange}, $L=12.26$, $\dx=L/63$, $\dt=5\cdot10^{-5}$,
512: $t\in[0,1000]$.
513: % N=63, dx=0.19466 L=63*0.19466=12.2636
514: % cut raws 0-500, t=10000 for 5000 raws, 2t.u. per raw
515: (d) Emergence of spontaneous periodic wavetrains via an instability of
516: periodic standing waves. Continuation of (c), $t\in[44460,45460]$.
517: % 7mx2-2230-2730cut, interval in file 40000-46000, 3000 raws
518: }{instab}
519: %%%%%%%%%%%%%%%%%%%%%%%%
520:
521: The next possible candidate is the instability of spatially uniform
522: oscillations with respect to perturbations which nonzero freqency and
523: nonzero wavenumber. This case is not considered in the CH
524: nomenclature, and is called ``Hopf'' instability of spatially homogeneous
525: oscillations, with the same variants as in the previous case, by RS.
526: This instability looks plausible as spatially homogeneous (spatially uniform) oscillations in our
527: systems are indeed possible and even stable in small spatial
528: domains, so we have investigated this possibility with particular
529: care. As limit cycles in the point systems of~\eq{TB} and~\eq{RM} can
530: not be described analytically, the investigation of stability has to
531: be done numerically. We have considered solutions of the form
532: $\u(x,t)=\uo(t)+\Re\left(\v(t)e^{i\k x}\right)$ with $|\v|\ll1$, which
533: gives a coupled system of ordinary differential equations
534: %
535: \begin{subequations}
536: \begin{align}
537: \Df{\uo}{t} &= \f(\uo), \label{coupled.non} \\
538: \Df{\v}{t} &= \left( \F(\uo) - \D(\uo) k^2 \right) \v , \label{coupled.lin}
539: \end{align} \label{coupled}
540: \end{subequations}
541: %
542: with parameter $\k$. We solved system~\eq{coupled} forward in time
543: with initial conditions for bulk oscillations $\uo$ in the basin of
544: attraction of the limit cycle, and arbitrary nonzero initial conditions
545: for the perturbation $\v$. Then we estimated the Lyapunov exponent for the
546: $\v$-subsystem, $\lyap(\k)=\lim_{t\to\infty}t^{-1}\ln(||\v(t)||)$.
547: The estimation was done by finding maxima of the first
548: component of $\v(t)$ and linearly fitting their logarithms against
549: $t$, for an interval of large enough values of $t$. For selected
550: values of $\k$ we used two linear independent sets of initial
551: conditions for $\v$, to eliminate the theoretical possibility of
552: accidentally chosing initial conditions that did not lead to the maximal
553: exponent.
554:
555:
556: The resulting graph $\lyap(\k)$ for the TB model at the same
557: parameters as in
558: \fig{density}(a) and \figref{lchange} is shown on \fig{instab}(a). For
559: comparison, we also show histograms of the empirical wavenumbers
560: observed in simulations shown in \fig{lchange}, calculated as $\k=2\pi
561: n/L$, separately for the growing and decreasing $L$.
562: \Fig{instab}(b) shows similar graphs made for the RM model
563: at the same parameters as in \fig{density}(f). It is clear that,
564: although there are finite bands of wavenumbers producing growing
565: perturbations, the
566: actually selected wavenumbers are not the same as those of
567: the fastest growing perturbations, and for the TB model they even partly fall in the interval of
568: decaying perturbations.
569:
570: Moreover, the growing perturbations of the spatially uniform oscillations in
571: fact do not represent propagating periodic waves, but standing
572: waves. This is illustrated in \fig{instab}(c) where we show a density
573: plot of a simulation of the full model, similar to \fig{density}(a)
574: but with different initial conditions. Here we chose initial
575: conditions as spatially uniform oscillations plus a very small perturbation
576: sinusoidal in space. Note that for the limit of infinitely small
577: perturbation amplitudes this exactly corresponds to system
578: \eq{coupled}.
579:
580: We conclude that although the cross-diffusion driven instability does indeed
581: take place in the considered examples, the waves that emerge are in
582: fact quite different from the spontaneous periodic traveling waves.
583:
584: \paragraph{Spontaneous sources as a precursor of spontaneous periodic waves}
585:
586: The periodic standing waves emerging via the cross-diffusion driven
587: instability described above, are in turn unstable themselves.
588: \Fig{instab}(d) shows a continuation of simulation of
589: \fig{instab}(c). The standing waves are observed for a long time, as
590: they are stable within the space of functions with spatial period
591: $2\pi/\k=L/6$, and the numerical initial conditions are almost exactly
592: periodic with that period, up to small errors resulting from finite
593: precision arithmetics. The small symmetry-breaking numerical errors
594: allow for an instability of the periodic standing waves to develop,
595: during which some of the standing waves occur later than others. When
596: this instability sufficiently develops, there is a sudden, ``hard''
597: transition to propagating waves. The spatial period of the propagating
598: waves is twice longer than the spatial period of preceding standing
599: waves. We stress that the traveling waves do not appear via anything
600: like ``bifurcation'' from standing waves, at least in the examples we
601: considered.
602:
603: Notice that the long transient solution shown in \fig{instab}(c,d) is
604: a periodic standing wave by its symmetry, but it also looks like a
605: periodic set of focal sources, synchronously sending out solitary
606: waves which then annihilate each other. As can be seen in
607: \fig{density}, apart from the symmetry, this sort of transient before
608: the onset of periodic waves is typical, and only its duration varies
609: in different simulations. That is, the special initial conditions in
610: \fig{instab}(c,d) only affect the symmetry and the duration of the
611: transient, but not its qualitative character. A similar route to
612: traveling waves via unstable periodic set of ``focal sources''
613: standing waves is obseved in the RM model.
614:
615: \section{Conclusion}
616:
617: The considered examples demonstrate an unusual type of behaviour. The
618: systems are oscillatory, but the spatially uniform oscillations are
619: unstable. The systems can also demonstrate standing periodic waves,
620: which are also unstable. These instabilities lead to periodic
621: propagating waves, which seems to be the only stable regime. This
622: regime emerges spontaneously even when boundary conditions disallow
623: propagating waves. The periods of the waves can be in a certain
624: interval with strict boundaries, both upper and lower. Nearer the
625: upper end of the interval, \ie\ at longer wavelengths, the periodic
626: waves do not propagate steadily but are modulated. Transition from
627: steady to modulated propagation is soft and has empirical features of
628: a supercritical Hopf bifurcation (of a relative equilibrium), \ie\
629: possibly an Eckhaus mechanism.
630:
631: The defining features described above are sufficiently generic, and
632: the phenomenon of spontaneous periodic traveling waves does not
633: disappear as the parameters are varied, nor it is restricted just to
634: one model. This behaviour does not fall into existing classification
635: of pattern formation scenarios. The detailed mechanisms
636: of spontaneous generation and maintenance of periodic traveling waves
637: require further investigation. However, it is clear that
638: cross-diffusion is an essential factor, since its replacement with,
639: or adding of significant amount of, self diffusion
640: eliminates the effect. Cross-diffusion phenomena are known in a in a
641: variety of physical situations. For example, spontaneous periodic
642: waves have been observed in a Burridge-Knopoff mathematical model of
643: earthquakes~\cite{Cartwright-etal-1997,Cartwright-etal-1999}. That
644: model belongs to the class \eq{RD}, with only one nonzero element
645: of matrix $\D$, as in our simulations shown in~\fig{density}(a) and
646: (g) but constant, and
647: excitable FitzHugh-Nagumo local kinetics. It is not known whether
648: the spontaneous waves in the Burridge-Knopoff model have a finite
649: interval of allowed wavenumbers, as illustrated by~\fig{lchange} for
650: our case, however other described features of those waves are
651: similar to those described here and are likely to have a
652: similar nature. Further investigation of the mechanism of generation
653: of such waves is a subject for further study which is of
654: broad physical interest as a new pattern forming mechanism in
655: dissipative spatially distributed systems.
656:
657: Returning to the application that originally motivated this study,
658: attempts to explain waves observed in cyclic biological populations,
659: using reaction-diffusion models, had to involve spatially-nonuniform
660: external factors, \eg\ sites of increased mortality due to
661: environmental conditions
662: \cite{Sherratt-etal-2002,Sherratt-Smith-2008}. Such factors are needed
663: to disallow uniform oscillations. Our present results imply that such
664: factors may not be necessary if cross-diffusion interaction are taken
665: into account as the uniform oscillations may be unstable and waves
666: form spontaneously.
667:
668: \textbf{Acknowledgments}
669: We are grateful to O.~Piro for helpful advice.
670: The study was supported in part by RFBR grant 07-04-00363 (Russia) and by a
671: grant from the Research Centre for Mathematics and Modelling of
672: Liverpool University (UK).
673:
674: %%%%%%%%%%%%%%%%%%%%
675: %\bibliography{opp}
676: \begin{thebibliography}{26}
677: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
678: \expandafter\ifx\csname bibnamefont\endcsname\relax
679: \def\bibnamefont#1{#1}\fi
680: \expandafter\ifx\csname bibfnamefont\endcsname\relax
681: \def\bibfnamefont#1{#1}\fi
682: \expandafter\ifx\csname citenamefont\endcsname\relax
683: \def\citenamefont#1{#1}\fi
684: \expandafter\ifx\csname url\endcsname\relax
685: \def\url#1{\texttt{#1}}\fi
686: \expandafter\ifx\csname urlprefix\endcsname\relax\def\urlprefix{URL }\fi
687: \providecommand{\bibinfo}[2]{#2}
688: \providecommand{\eprint}[2][]{\url{#2}}
689:
690: \bibitem[{\citenamefont{Cross and Hohenberg}(1993)}]{Cross-Hohenberg-1993}
691: \bibinfo{author}{\bibfnamefont{M.~C.} \bibnamefont{Cross}} \bibnamefont{and}
692: \bibinfo{author}{\bibfnamefont{P.~C.} \bibnamefont{Hohenberg}},
693: \bibinfo{journal}{Rev. Mod. Phys.} \textbf{\bibinfo{volume}{65}},
694: \bibinfo{pages}{851} (\bibinfo{year}{1993}).
695:
696: \bibitem[{\citenamefont{Tsyganov et~al.}(2003)\citenamefont{Tsyganov, Brindley,
697: Holden, and Biktashev}}]{QS1}
698: \bibinfo{author}{\bibfnamefont{M.~A.} \bibnamefont{Tsyganov}},
699: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Brindley}},
700: \bibinfo{author}{\bibfnamefont{A.~V.} \bibnamefont{Holden}},
701: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{V.~N.}
702: \bibnamefont{Biktashev}}, \bibinfo{journal}{Phys. Rev. Lett.}
703: \textbf{\bibinfo{volume}{91}}, \bibinfo{pages}{218102}
704: (\bibinfo{year}{2003}).
705:
706: \bibitem[{\citenamefont{Tsyganov et~al.}(2004)\citenamefont{Tsyganov, Brindley,
707: Holden, and Biktashev}}]{QS2}
708: \bibinfo{author}{\bibfnamefont{M.~A.} \bibnamefont{Tsyganov}},
709: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Brindley}},
710: \bibinfo{author}{\bibfnamefont{A.~V.} \bibnamefont{Holden}},
711: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{V.~N.}
712: \bibnamefont{Biktashev}}, \bibinfo{journal}{Physica D}
713: \textbf{\bibinfo{volume}{197}}, \bibinfo{pages}{18} (\bibinfo{year}{2004}).
714:
715: \bibitem[{\citenamefont{Biktashev et~al.}(2004)\citenamefont{Biktashev,
716: Brindley, Holden, and Tsyganov}}]{QS3}
717: \bibinfo{author}{\bibfnamefont{V.~N.} \bibnamefont{Biktashev}},
718: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Brindley}},
719: \bibinfo{author}{\bibfnamefont{A.~V.} \bibnamefont{Holden}},
720: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{M.~A.}
721: \bibnamefont{Tsyganov}}, \bibinfo{journal}{Chaos}
722: \textbf{\bibinfo{volume}{14}}, \bibinfo{pages}{988} (\bibinfo{year}{2004}).
723:
724: \bibitem[{\citenamefont{Tsyganov and Biktashev}(2004)}]{QS5}
725: \bibinfo{author}{\bibfnamefont{M.~A.} \bibnamefont{Tsyganov}} \bibnamefont{and}
726: \bibinfo{author}{\bibfnamefont{V.~N.} \bibnamefont{Biktashev}},
727: \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{70}},
728: \bibinfo{pages}{031901} (\bibinfo{year}{2004}).
729:
730: \bibitem[{\citenamefont{Biktashev and Tsyganov}(2005)}]{QS6}
731: \bibinfo{author}{\bibfnamefont{V.~N.} \bibnamefont{Biktashev}}
732: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{M.~A.}
733: \bibnamefont{Tsyganov}}, \bibinfo{journal}{Proc. Roy. Soc. Lond. A}
734: \textbf{\bibinfo{volume}{461}}, \bibinfo{pages}{3711} (\bibinfo{year}{2005}).
735:
736: \bibitem[{\citenamefont{Tsyganov et~al.}(2007)\citenamefont{Tsyganov,
737: Biktashev, Brindley, Holden, and Ivanitsky}}]{UFN07}
738: \bibinfo{author}{\bibfnamefont{M.~A.} \bibnamefont{Tsyganov}},
739: \bibinfo{author}{\bibfnamefont{V.~N.} \bibnamefont{Biktashev}},
740: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Brindley}},
741: \bibinfo{author}{\bibfnamefont{A.~V.} \bibnamefont{Holden}},
742: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{G.~R.}
743: \bibnamefont{Ivanitsky}}, \bibinfo{journal}{Physics-Uspekhi}
744: \textbf{\bibinfo{volume}{50}}, \bibinfo{pages}{275} (\bibinfo{year}{2007}).
745:
746: \bibitem[{\citenamefont{Vanag and Epstein}(2009)}]{Vanag-Epstein-2009}
747: \bibinfo{author}{\bibfnamefont{V.~K.} \bibnamefont{Vanag}} \bibnamefont{and}
748: \bibinfo{author}{\bibfnamefont{I.~R.} \bibnamefont{Epstein}},
749: \bibinfo{journal}{Phys. Chem. Chem. Phys.} \textbf{\bibinfo{volume}{11}},
750: \bibinfo{pages}{897} (\bibinfo{year}{2009}).
751:
752: \bibitem[{\citenamefont{del Castillo~Negrete et~al.}(2002)\citenamefont{del
753: Castillo~Negrete, Carreras, and Lynch}}]{del-Castillo-Negrete-etal-2002}
754: \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{del Castillo~Negrete}},
755: \bibinfo{author}{\bibfnamefont{B.~A.} \bibnamefont{Carreras}},
756: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{V.}~\bibnamefont{Lynch}},
757: \bibinfo{journal}{Physica D} \textbf{\bibinfo{volume}{168--169}},
758: \bibinfo{pages}{45} (\bibinfo{year}{2002}).
759:
760: \bibitem[{\citenamefont{Chung and
761: Peacock-L{\'o}pez}(2007)}]{Chung-Peacock-2007}
762: \bibinfo{author}{\bibfnamefont{J.~M.} \bibnamefont{Chung}} \bibnamefont{and}
763: \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Peacock-L{\'o}pez}},
764: \bibinfo{journal}{Phys. Lett. A} \textbf{\bibinfo{volume}{371}},
765: \bibinfo{pages}{41} (\bibinfo{year}{2007}).
766:
767: \bibitem[{\citenamefont{Truscott and Brindley}(1994)}]{Truscott-Brindley-1994}
768: \bibinfo{author}{\bibfnamefont{J.~E.} \bibnamefont{Truscott}} \bibnamefont{and}
769: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Brindley}},
770: \bibinfo{journal}{Phil. Trans. Roy. Soc. Lond. ser. A}
771: \textbf{\bibinfo{volume}{347}}, \bibinfo{pages}{703} (\bibinfo{year}{1994}).
772:
773: \bibitem[{\citenamefont{Murray}(2002)}]{Murray-2002}
774: \bibinfo{author}{\bibfnamefont{J.~D.} \bibnamefont{Murray}},
775: \emph{\bibinfo{title}{Mathematical Biology I: An Introduction}}
776: (\bibinfo{publisher}{Springer}, \bibinfo{address}{New York, NY},
777: \bibinfo{year}{2002}).
778:
779: \bibitem[{\citenamefont{Britton}(2003)}]{Britton-2003}
780: \bibinfo{author}{\bibfnamefont{N.~F.} \bibnamefont{Britton}},
781: \emph{\bibinfo{title}{Essential mathematical biology}}
782: (\bibinfo{publisher}{Springer}, \bibinfo{address}{New York, NY},
783: \bibinfo{year}{2003}).
784:
785: \bibitem[{\citenamefont{Sherratt and Smith}(2008)}]{Sherratt-Smith-2008}
786: \bibinfo{author}{\bibfnamefont{J.~A.} \bibnamefont{Sherratt}} \bibnamefont{and}
787: \bibinfo{author}{\bibfnamefont{M.~J.} \bibnamefont{Smith}},
788: \bibinfo{journal}{Journal of the Royal Society Interface}
789: \textbf{\bibinfo{volume}{5}}, \bibinfo{pages}{483} (\bibinfo{year}{2008}).
790:
791: \bibitem[{\citenamefont{Rosenzweig and
792: MacArthur}(1963)}]{Rosenzweig-MacArthur-1963}
793: \bibinfo{author}{\bibfnamefont{M.~L.} \bibnamefont{Rosenzweig}}
794: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{R.~H.}
795: \bibnamefont{MacArthur}}, \bibinfo{journal}{American Naturalist}
796: \textbf{\bibinfo{volume}{97}}, \bibinfo{pages}{209} (\bibinfo{year}{1963}).
797:
798: \bibitem[{\citenamefont{Rademacher and Scheel}(2007)}]{Rademacher-Scheel-2007}
799: \bibinfo{author}{\bibfnamefont{J.~D.~N.} \bibnamefont{Rademacher}}
800: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Scheel}},
801: \bibinfo{journal}{Int. J. of Bifurcation and Chaos}
802: \textbf{\bibinfo{volume}{17}}, \bibinfo{pages}{2679} (\bibinfo{year}{2007}).
803:
804: \bibitem[{\citenamefont{Kopell and Howard}(1973)}]{Kopell-Howard-1973}
805: \bibinfo{author}{\bibfnamefont{N.}~\bibnamefont{Kopell}} \bibnamefont{and}
806: \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Howard}},
807: \bibinfo{journal}{Stud. Appl. Math.} \textbf{\bibinfo{volume}{52}},
808: \bibinfo{pages}{291} (\bibinfo{year}{1973}).
809:
810: \bibitem[{\citenamefont{Swift and Hohenberg}(1977)}]{Swift-Hohenberg-1977}
811: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Swift}} \bibnamefont{and}
812: \bibinfo{author}{\bibfnamefont{P.~C.} \bibnamefont{Hohenberg}},
813: \bibinfo{journal}{Phys. Rev. A} \textbf{\bibinfo{volume}{15}},
814: \bibinfo{pages}{319} (\bibinfo{year}{1977}).
815:
816: \bibitem[{\citenamefont{Haken}(1983)}]{Haken-1983}
817: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Haken}},
818: \emph{\bibinfo{title}{Synergetics}} (\bibinfo{publisher}{Springer},
819: \bibinfo{address}{Berlin, Heidelberg, New York, Tokyo},
820: \bibinfo{year}{1983}).
821:
822: \bibitem[{\citenamefont{Livshits}(1983)}]{Livshits-1983}
823: \bibinfo{author}{\bibfnamefont{M.~A.} \bibnamefont{Livshits}},
824: \bibinfo{journal}{Z. Phys. B} \textbf{\bibinfo{volume}{53}},
825: \bibinfo{pages}{83} (\bibinfo{year}{1983}).
826:
827: \bibitem[{\citenamefont{Lega et~al.}(1994)\citenamefont{Lega, Moloney, and
828: Newell}}]{Lega-etal-1994}
829: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Lega}},
830: \bibinfo{author}{\bibfnamefont{J.~V.} \bibnamefont{Moloney}},
831: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{A.~C.}
832: \bibnamefont{Newell}}, \bibinfo{journal}{Phys. Rev. Lett.}
833: \textbf{\bibinfo{volume}{73}}, \bibinfo{pages}{2978} (\bibinfo{year}{1994}).
834:
835: \bibitem[{\citenamefont{Mendelson and Lega}(1998)}]{Mendelson-Lega-1998}
836: \bibinfo{author}{\bibfnamefont{N.~H.} \bibnamefont{Mendelson}}
837: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Lega}},
838: \bibinfo{journal}{Journal of Bacteriology} \textbf{\bibinfo{volume}{180}},
839: \bibinfo{pages}{3285} (\bibinfo{year}{1998}).
840:
841: \bibitem[{\citenamefont{Vanag and Epstein}(2002)}]{Vanag-Epstein-2002}
842: \bibinfo{author}{\bibfnamefont{V.~K.} \bibnamefont{Vanag}} \bibnamefont{and}
843: \bibinfo{author}{\bibfnamefont{I.~R.} \bibnamefont{Epstein}},
844: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{88}},
845: \bibinfo{pages}{088303} (\bibinfo{year}{2002}).
846:
847: \bibitem[{\citenamefont{Cartwright et~al.}(1997)\citenamefont{Cartwright,
848: Hernandez-Garcia, and Piro}}]{Cartwright-etal-1997}
849: \bibinfo{author}{\bibfnamefont{J.~H.~E.} \bibnamefont{Cartwright}},
850: \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Hernandez-Garcia}},
851: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{O.}~\bibnamefont{Piro}},
852: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{79}},
853: \bibinfo{pages}{527} (\bibinfo{year}{1997}).
854:
855: \bibitem[{\citenamefont{Cartwright et~al.}(1999)\citenamefont{Cartwright,
856: Eguiliz, Hernandez-Garcia, and Piro}}]{Cartwright-etal-1999}
857: \bibinfo{author}{\bibfnamefont{J.~H.~E.} \bibnamefont{Cartwright}},
858: \bibinfo{author}{\bibfnamefont{V.~M.} \bibnamefont{Eguiliz}},
859: \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Hernandez-Garcia}},
860: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{O.}~\bibnamefont{Piro}},
861: \bibinfo{journal}{Int. J. of Bifurcation and Chaos}
862: \textbf{\bibinfo{volume}{9}}, \bibinfo{pages}{2197} (\bibinfo{year}{1999}).
863:
864: \bibitem[{\citenamefont{Sherratt et~al.}(2002)\citenamefont{Sherratt, Lambin,
865: Thomas, and Sherratt}}]{Sherratt-etal-2002}
866: \bibinfo{author}{\bibfnamefont{J.~A.} \bibnamefont{Sherratt}},
867: \bibinfo{author}{\bibfnamefont{X.}~\bibnamefont{Lambin}},
868: \bibinfo{author}{\bibfnamefont{C.~J.} \bibnamefont{Thomas}},
869: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{T.~N.}
870: \bibnamefont{Sherratt}}, \bibinfo{journal}{Proc. Roy. Soc. Lond. ser. B}
871: \textbf{\bibinfo{volume}{269}}, \bibinfo{pages}{327} (\bibinfo{year}{2002}).
872:
873: \end{thebibliography}
874:
875:
876: \end{document}
877: